Sample size for canonical correlation analysis in corn

The canonical correlation analysis has been successfully used in many areas aiming to extract important information from a pair of data sets. Thus, the objective this work was determine sample size (number plants) required estimate correlations corn characteristics. Six characteristics were measured 361, 373, and 416 plants, respectively, single, three-way double cross hybrids 2008/2009 crop year 1,777, 1,693, 1,720 three-way, (2009/2010 crop) (six cases). analyses carried out between group plant architecture (plant height at harvest ear insertion height) versus grain production (hundred grains mass per plant) (scenario 1), dimensions (ear length diameter) 2). for estimation determined by resampling with replacement application model linear response plateau. Measuring 270 plants is sufficient groups two each corn. This can be as reference reliable analysis.